HLA localization of the genes predisposing to

Brief Genetics Report Non Class II HLA Gene Associated With Type 1 Diabetes Maps to the 240-kb Region Near HLA-B Sergei Nejentsev, Zsofia Gombos, Antti-Pekka Laine, Riitta Veijola, Mikael Knip, Olli Simell, Outi Vaarala, Hans K. Åkerblom, and Jorma Ilonen Several studies provide evidence that in addition to the DQ-DR genes, HLA contains another uncharacterized gene or genes associated with type 1 diabetes. Our aim was to investigate the effect of this gene independently of the DQ-DR genes and to localize it with a matched case-control study. More than 1,400 patients and 30,000 control individuals from Finland were studied. They were first genotyped for the selected alleles of the HLA-DQB1, -DQA1, and -DRB1 genes. For the DR3/4(0404) genotype, 75 patients and 181 control subjects were stratified, and 241 patients and 354 controls were stratified for the DR3/4(0401) genotype. Ten microsatellite markers in the HLA class III and I regions (D6S273, TNFa, C12A, STR MICA, MIB, C125, C143, C245, C3211, and MOGc) and selected alleles of the HLA-A and HLA-B genes were studied. In the DR3/4(0404)-stratified group, we found that markers located between C12A and C143 near the HLA-B gene confer a strong additional diabetes association. This was confirmed by the population differentiation test in both DR3/4(0404)- and DR3/4(0401)-stratified groups. Our data indicate that an additional gene associated with type 1 diabetes is located in the 240-kb region near HLA-B. We excluded STR MICA polymorphism as a mutation responsible for diabetes association. Diabetes 49:2217 2221, 2000 HLA localization of the genes predisposing to type 1 diabetes is well established. DQB1, DQA1, and DRB1 genes are considered primary and major loci associated with type 1 diabetes. However, evidence is accumulating that the HLA class I III region contains gene(s) associated with an increased risk From the Juvenile Diabetes Foundation Center for Type 1 Diabetes Prevention in Finland (S.N., Z.G., A.-P.L., R.V., M.K., O.S., J.I.); the Department of Virology (S.N., Z.G., A.-P.L., J.I.), University of Turku, Turku; the Department of Pediatrics (R.V., M.K.), University of Oulu, Oulu; the Department of Pediatrics (M.K.), Medical School, University of Tampere, Tampere; the Department of Pediatrics (O.S.), University of Turku, Turku; the National Public Health Institute (O.V.), Helsinki; and the Hospital for Children and Adolescents (H.K.Å.), University of Helsinki, Helsinki, Finland. Address correspondence and reprint requests to Sergei Nejentsev, MD, PhD, Department of Virology, University of Turku, Kiinamyllynkatu 13, FIN-20520 Turku, Finland. E-mail: serneje@utu.fi. Received for publication 27 March 2000 and accepted in revised form 17 August 2000. Abbreviations: OR, odds ratio; PCR, polymerase chain reaction. for type 1 diabetes (1 4) or its earlier age at onset (5 8). We searched for an additional HLA gene associated with type 1 diabetes among Finnish patients and control subjects stratified for the DQ-DR alleles. Initially, samples from >1,400 type 1 diabetic patients and >30,000 control subjects were collected and genotyped for DQB1 (9). DQB1*02/0302 subjects were then studied for the DQA1*05 and *0201 alleles and DR4 subtypes (of which only DRB1*0401 and DRB1*0404 are frequent among Finns). Thus, a group stratified for the DQA1*05-DQB1*02/ DQB1*0302-DRB1*0404 [DR3/4(0404)] genotype (75 patients and 181 control subjects) and a group stratified for the DQA1*05-DQB1*02/DQB1*0302-DRB1*0401 [DR3/4(0401)] genotype (241 patients and 354 control subjects) were selected. These subjects are all DR3/DR4 haplotype heterozygotes: one group carries DR404 and the other DR401 haplotype. Thus, within each group we excluded the diabetogenic effect of the known DQ-DR markers. In these groups, selected alleles of the HLA-B and HLA-A genes and 10 microsatellites (D6S273, TNFa, C12A, STR MICA, MIB, C125, C143, C245, C3211, and MOGc) were studied to map a non class II HLA gene associated with type 1 diabetes susceptibility. These cover distance of ~2,300 kb and include the whole HLA class I region and the class I III boundary (Fig. 1). We found that in the DR3/404- and DR3/401-stratified groups there was only one allele at each microsatellite locus present in the majority of both patients and control subjects that did not confer any risk or protection (Table 1). Because DR3 haplotype is present in both DR3/404- and DR3/401- stratified groups, these alleles were supposed to represent the extended DR3 haplotype: DQB1*02- DQA1*05- D6S273*141-TNFa*2-C12A*256-STR MICA*A5.1-MIB*350- C125*194-C143*449-C245*436-C3211*217. This was confirmed by the study of parents of 20 randomly selected DR3/404 + and 20 DR3/401 + type 1 diabetic patients, in which the direct study of haplotypes was possible (data not shown). This finding implies that in the present study, the DR3 haplotype does not contain any additional diseaseassociated marker and could be explained by high homogeneity of the DR3 haplotype in the Finnish population. In the DR3/404-stratified group, we confirmed strong diabetes association of B39, previously shown by us (3,4), and of A24, shown by others (7,8). There was only one allele at each microsatellite locus associated with an increased risk for DIABETES, VOL. 49, DECEMBER 2000 2217

NEW TYPE 1 DIABETES GENE MAPS NEAR HLA-B FIG. 1. Markers in the HLA region. Approximate distances (kb) between markers are shown above., genes;, microsatellites. type 1 diabetes, and all of these alleles, as well as A24, were strongly associated with B39 (Table 1). None of the other alleles of the microsatellite markers studied were found to be associated with B39 (data not shown), and none were significantly increased among patients (Table 1). Based on these data and haplotype analysis in parents of 20 patients with the DR3/404 genotype selected randomly, we reconstructed a haplotype, DQB1*0302-DRB1*0404-D6S273*137-TNFa*6- C12A*250-STR MICA*A9-MIB*332-B39-C125*200-C143*425- C245*436-C3211*207-A24-MOGc*130. This DR404 haplotype is associated with an increased risk for type 1 diabetes, and disease risk conferred by the DQ-DR segment of this haplotype is increased by alleles of the D6S273-MOGc segment. Two tests were applied to map the region of the strongest type 1 diabetes association within the D6S273-MOGc segment, which should contain the disease-predisposing gene. First, odds ratios (ORs) for the strongest associated allele of the above-described haplotype were compared (Table 1 and Fig. 2A). None of the markers were associated with higher risk than the HLA-B39 allele. Despite the fact that the 95% CIs were overlapping in all comparisons, associated alleles in the region between the C12A and C143 markers conferred higher risk than the alleles of the other markers, which suggests that the type 1 diabetes associated gene is located in this area. Statistical evidence for this hypothesis was obtained by the population differentiation test. Thus, comparison of allele distribution in the DR3/404-stratified group reveals that the patient population is different from control population at the C12A, STR MICA, C125, and C143 markers, again emphasizing the importance of the region between the C12A and C143 loci (Table 2 and Fig. 2B). In the DR3/401-stratified group of patients and control subjects, no single allele was associated with the significant increase or decrease of risk conferred by the DQ-DR markers (Fig. 3A). However, results of the population differentiation test in the DR3/401-stratified group demonstrate that the population of patients is different from the background population by allele distribution at the MIB locus. The test result is also very close to the significance threshold at the C12A locus (Table 2 and Fig. 3B). It is remarkable that these loci are located within the region suggested as candidate in the study of the DR3/404-stratified group. The present study does not support an idea that the polymorphism of the transmembrane region of the MICA protein (STR MICA) could itself be responsible for diabetes association. Indeed, in the DR3/404-stratified group, STR MICA*A9 confers high risk for type 1 diabetes (OR 4.0, 95% CI 2.1 7.6), whereas in the DR3/401-stratified group the same polymorphism confers significantly ( 2 = 12.0, 1 df, P < 0.0006) lower risk (OR 1.03, 95% CI 0.6 1.7). Earlier studies demonstrate that the type 1 diabetes gene in the HLA class I III regions is associated with an earlier age at onset (5 8). Hence, it is important to show that the putative polymorphism, which we map near HLA-B, confers not only an increased risk for type 1 diabetes, but also an earlier age at onset. This was tested in the group of DR3/404-stratified type 1 diabetic patients. The frequency of B39 among patients that were diagnosed at 10 years of age or earlier was 19 of 40 (47.5%), whereas it was 7 of 22 (31.8%) among those diagnosed at 11 years of age or later (P = 0.18, OR = 1.9). The difference was significant in the group of patients diagnosed between 5 and 9 years of age (9 of 13 [69.2%] vs. 17 of 49 [34.7%] diagnosed in the other age groups [P < 0.03, OR 4.2]). Interestingly, in the study of Japanese patients, A24 was mostly increased in the group of patients diagnosed between the ages of 6 and 12 years but not in the youngest age group (8). In the present study, non class II HLA markers of the DR404 haplotype, D6S273*137-TNFa*6-C12A*250- STR MICA*A9-MIB*332-B39-C125*200-C143*425-C245*436- C3211*207-A24-MOGc*130, had similar but lower effects on the age at onset than B39 (data not shown). Thus, we suggest that association of markers in HLA class I III regions with young age at onset of type 1 diabetes is a consequence of linkage disequilibrium between these markers and the putative polymorphism located in the region near HLA-B. Several attempts to map a non class II HLA gene associated with type 1 diabetes have been done so far. At first, by analyzing the DR3-B8 and DR3-B18 haplotypes, it was predicted that the region between HLA-B and BAT3 should contain a gene associated with type 1 diabetes (10). This study is in line with our results, because suggested regions overlap between C12A and HLA-B (Fig. 1). Another study suggests that the non class II HLA gene of type 1 diabetes susceptibility is located between D6S273 and TNFa markers (11). However, microsatellite D6S265, which was used as a marker next to TNFa is located ~1.3 Mb telomerically. Because no markers were tested between these two markers, to our mind, the telomeric boundary of the candidate region could not be specified exactly. Finally, a non class II HLA gene associated with 2218 DIABETES, VOL. 49, DECEMBER 2000

S. NEJENTSEV AND ASSOCIATES TABLE 1 Frequency (%) of the HLA markers among type 1 diabetic patients and control subjects stratified for the DR3/404 genotype Patients Control subjects Marker*allele (n = 75) (n = 181) OR 95% CI P P cor D6S273*135 25.3 42.5 0.5 0.2 0.9 0.01 NS *137 74.7 58.0 2.1 1.1 4.1 0.02 NS *141 86.7 76.2 2.0 0.9 4.6 NS TNFa*2 88.0 82.9 1.5 0.6 3.6 NS *6 50.7 34.3 2.0 1.1 3.5 0.02 NS *11 38.7 48.1 0.7 0.4 1.2 NS C12A*236 12.0 28.7 0.3 0.15 0.8 0.007 NS *250 49.3 19.3 4.1 2.2 7.6 0.000002 <0.0003 *256 81.3 76.2 1.4 0.7 2.8 NS MICA*A4 12.0 23.2 0.5 0.2 1.0 NS *A5.1 93.3 93.4 1.0 0.3 3.4 NS *A9 45.3 17.1 4.0 2.1 7.6 0.000005 <0.0007 MIB*326 10.7 17.7 0.6 0.2 1.3 NS *332 41.3 15.5 3.8 2.0 7.4 0.00002 0.002 *350 86.7 90.6 0.7 0.3 1.7 NS HLA-B*39 41.3 12.7 4.8 2.5 9.6 0.000001 <0.0002 C125*194 78.7 74.6 1.3 0.6 2.5 NS *200 41.3 14.9 4.0 2.1 7.8 0.000009 <0.002 *210 29.3 33.7 0.8 0.4 1.5 NS C143*425 37.3 12.7 4.1 2.1 8.2 <0.00002 <0.002 *445 37.3 49.2 0.6 0.3 1.1 NS *449 72.0 70.2 1.1 0.6 2.1 NS C245*428 24.0 26.5 0.9 0.4 1.7 NS *436 90.7 80.7 2.3 0.93 6.1 NS *476 20.0 33.7 0.5 0.2 1.0 0.04 NS *436/436 36.0 21.0 2.1 1.1 4.0 <0.02 NS C3211*207 49.3 23.8 3.1 1.7 5.7 0.0001 <0.02 *211 21.3 35.4 0.5 0.3 1.0 0.04 NS *217 58.7 61.3 0.9 0.5 1.6 NS HLA-A*24 38.7 19.3 2.6 1.4 5.0 <0.002 NS MOGc*130 66.7 51.9 1.9 1.0 3.4 0.04 NS *132 42.7 49.2 0.8 0.4 1.4 NS *148 42.7 41.4 1.1 0.6 1.9 NS Generally referred allele names were used for the TNFa and STR MICA microsatellites. For the other microsatellite loci, allele number corresponds to the length of the PCR product (exactly estimated by sequencing of homozygous samples). Up to three of the most common alleles of each microsatellite marker are shown. P cor, P corrected (P 123). Allele associated with the DR404-B39 haplotype (P < 0.0001 for association with B39); alleles associated with the DR3 haplotype; alleles associated with the DR404-B39 haplotype (P < 0.00000001 for association with B39). type 1 diabetes was mapped near D6S2223 (12), ~2.5 Mb telomeric to MOGc. Our data demonstrate that the effect seen by us is already weak at MOGc and hardly extend to D6S2223. The most probable explanation of these findings is the existence of at least three loci of type 1 diabetes susceptibility in HLA. One of them is located at the HLA class II region and consists of the DQB1-DQA1-DRB1 genes. Another maps to the centromeric end of the class I region near HLA-B, whereas the third gene is located at the region flanking the HLA complex, telomeric to the class I region. This hypothesis is supported by structure of the Idd1 locus in the NOD mice (13). The 240-kb region mapped in the present study between C12A and C143 includes the well-known HLA-B, HLA-C, MICA, and MICB genes, as well as less-characterized P5-1 and 3.8-1.1 genes (14). Markers located in this region were associated with several autoimmune diseases (15 17). Future studies of the variants affecting protein sequence or expression of the genes in the mapped 240-kb region will identify the responsible polymorphism(s) associated with type 1 diabetes and other autoimmune diseases. RESEARCH DESIGN AND METHODS Subjects. Type 1 diabetic patients were recruited in Finland (mostly in Turku, Oulu, Helsinki, and Tampere) and were diagnosed before the age of 16 years. Samples from newborns were collected in Turku, Oulu, and Tampere as part of the Diabetes Prediction and Prevention project and were used as control subjects. All control subjects were tested for islet cell antibodies and positives were excluded from the analysis. The study was approved by the ethical committees of the participating hospitals. Informed consent was obtained from the participating subjects and/or their parents. Genotyping. Samples were collected as EDTA blood, and dried blood spots of 50 µl were prepared for screening of six DQB1 alleles, including *02 and *0302 (9), DR4 subtypes and DQA1*05 and *0201 alleles. The DR3/4-stratified groups were selected because this genotype is the most common among the Finnish type 1 diabetic patients and, hence, more class II-stratified patients were available. DNA was extracted using a Quiagen DNA extraction kit from the EDTA blood of the DR3/4 + subjects. The DR404 haplotype was studied, because our previous observations demonstrate that the HLA-B39 allele was DIABETES, VOL. 49, DECEMBER 2000 2219

NEW TYPE 1 DIABETES GENE MAPS NEAR HLA-B TABLE 2 Population differentiation test for comparison of type 1 diabetic patients (n = 75) and control subjects (n = 181), stratified for the DR3/404 genotype, as well as type 1 diabetic patients (n = 241) and control subjects (n = 354), stratified for the DR3/401 genotype DR3/404-stratified DR3/401-stratified group group Locus P SE P SE D6S273 0.02368 0.00040 0.09557 0.00114 TNFa 0.29543 0.00166 0.65175 0.00178 C12A 0.00326 0.00013 0.01373 0.00041 STR MICA 0.00018 0.00002 0.54007 0.00195 MIB 0.02614 0.00047 0.00463 0.00020 C125 0.00239 0.00013 0.56609 0.00229 C143 0.00013 0.00002 0.46173 0.00242 C245 0.16287 0.00117 0.27557 0.00239 C3211 0.02037 0.00042 0.14022 0.00189 MOGc 0.53826 0.00163 0.25469 0.00236 Significant P values of the population differentiation test (assuming P < 0.01 threshold) are bold. SE, standard error. small sample size (20). Values of P < 0.01 were accepted as a significance threshold for rejecting the null hypothesis of the allelic distribution identity between the population of type 1 diabetic patients and the control population. Because genotyping of HLA-A and HLA-B genes was not full-house, only microsatellite loci were analyzed by the population differentiation test. FIG. 2. ORs for the strongest associated allele (A) and the population differentiation test (B) in the DR3/404-stratified group., OR;, 95% CI for OR;, P value for population differentiation test;, significance threshold. associated with a conspicuously high risk on this haplotype (3,4). At the same time, it was shown that B39 is in linkage disequilibrium with the HLA-A24 allele on the DR404 haplotype (4). Our previous studies failed to identify any HLA-B alleles that could modify the disease risk in the DR401 haplotype. However, the study of this haplotype also seems important, both for confirmation and fine mapping through the overlapping of the candidate regions on both haplotypes. B62 and A2, the most frequent alleles of class I genes in linkage disequilibrium with DR401 in the Finnish population (4), were also studied. So, the prevalence of A24 and B39 among DR3/404 + subjects and of A2 and B62 among DR3/401 + subjects was tested by means of the sequence-specific polymerase chain reaction (PCR) (18). Microsatellite markers were studied using PCR primers (19), labeled with fluorescent dyes, ABI 310 electrophoresis, and the Genescan software. Statistical analysis. Allele frequencies were estimated by direct counting. 2 test or Fisher s exact test when appropriate were used for the calculation of statistical significance, and the level was set at P < 0.05. Bonferroni correction was done to correct P values for multiple comparisons; P cor is then presented. ORs were calculated according to the Woolf s formula. The Mantel-Haenszel procedure was used to test homogeneity of ORs. The Epistat program was used for calculations of these, as well as 95% CI. Population differentiation (genic) test was performed by the Genepop software package (20) with the following initial parameters: 1,000 steps of dememorization and 1,000 batches with 10,000 iterations per batch. Population differentiation test is an exact test that has an advantage of not being biased by rare alleles and ACKNOWLEDGMENTS This study was supported by grants from the Academy of Finland, the Emil Aaltonen Foundation, Turku University Central Hospital, and Juvenile Diabetes Foundation International (Grant 4 1998 274). Z.G. was supported by the scholarship of the European Community EURODIAB TRIGER project. S. N. was a PhD student of the Turku Graduate School for Biomedical Sciences, Turku, Finland. We thank Terttu Laurén, Ritva Suominen, and Mia Karlson for their skillful technical assistance. We also thank Drs. H. Fagerström, M. Flittner, B. Gustafsson, L. Herva, P. Hiltunen, T. Huupponen, M. Hyttinen, E.A. Kaprio, J. Komulainen, P. Korpela, E. Korpinen, A. Kuusisto, M-L. Käär, L. Laine, J. Lappalainen, P. Lautala, M. Lipsanen-Nyman, J. Mäenpää, K. Niemi, A. Nuuja, P. Ojajärvi, A. Putto-Laurila, P. Piekkola, S. Pöntynen, J. Sankala, I. Sipilä, J.-E. Sunell, V. Tommiska, T. Uotila, M. Väre, and P. Varimo for their assistance in collecting samples for this study and Prof. John Todd for his valuable comments. REFERENCES 1. Thomsen M, Molvig J, Zerbib A, de Preval C, Abbal M, Dugoujon JM, Ohayon E, Svejgaard A, Cambon Thomsen A, Nerup J: The susceptibility to insulindependent diabetes mellitus is associated with C4 allotypes independently of the association with HLA-DQ alleles in HLA-DR3,4 heterozygotes. Immunogenetics 28:320 327, 1988 2. Hanifi Moghaddam PH, Zwinderman AH, de Knijff P, Roep BO, Schipper RF, Van der Auwera B, Naipal A, Gorus F, Schuit F, Giphart MJ, the Belgian Diabetes Registry: TNFa microsatellite polymorphism modulates the risk of IDDM in Caucasians with the high-risk genotype HLA DQA1*0501- DQB1*0201/DQA1*0301-DQB1*0302. Diabetes 46:1514 1515, 1997 3. Nejentsev S, Reijonen H, Adojaan B, Kovalchuk L, Sochnevs A, Schwartz EI, Åkerblom HK, Ilonen J: The effect of HLA-B allele on the IDDM risk defined by DRB1*04 subtypes and DQB1*0302. Diabetes 46:1888 1892, 1997 2220 DIABETES, VOL. 49, DECEMBER 2000

S. NEJENTSEV AND ASSOCIATES FIG. 3. ORs for the strongest associated allele (A) and the population differentiation test (B) in the DR3/401-stratified group., OR;, 95% CI for OR;, P value for population differentiation test;, significance threshold. 4. Reijonen H, Nejentsev S, Tuokko J, Koskinen S, Tuomilehto-Wolf E, Åkerblom HK, Ilonen J: HLA-DR4 subtype and -B alleles in DQB1*0302-positive haplotypes associated with IDDM. Eur J Immunogenet 24:357 363, 1997 5. Fujisawa T, Ikegami H, Kawaguchi Y, Yamato E, Takekawa K, Nakagawa Y, Hamada Y, Ueda H, Shima K, Ogihara T: Class I HLA is associated with ageat-onset of IDDM, while class II HLA confers susceptibility to IDDM. Diabetologia 38:1493 1495, 1995 6. Demaine AG, Hibberd ML, Mangles D, Millward BA: A new marker in the HLA class I region is associated with the age at onset of IDDM. Diabetologia 38:623 628, 1995 7. Honeyman MC, Harrison LC, Drummond B, Colman PG, Tait BD: Analysis of families at risk for insulin-dependent diabetes mellitus reveals that HLA antigens influence progression to clinical disease. Mol Med 1:576 582, 1995 8. Mizota M, Uchigata Y, Moriyama S, Tokunaga K, Matsuura N, Miura J, Juji T, Omori Y: Age-dependent association of HLA-A24 in Japanese IDDM patients. Diabetologia 39:371 373, 1996 9. Sjöroos M, Iitiä A, Ilonen J, Reijonen H, Lövgren T: Triple-label hybridization assay for type-1 diabetes-related HLA alleles. Biotechniques 18:870 877, 1995 10. Degli Esposti MA, Abraham LJ, McCann V, Spies T, Christiansen FT, Dawkins RL: Ancestral haplotypes reveal the role of the central MHC in the immunogenetics of IDDM. Immunogenetics 36:345 356, 1992 11. Hanifi Moghaddam P, de Knijf P, Roep BO, Van der Auwera B, Naipal A, Gorus F, Schuit F, Giphart MJ, the Belgian Diabetes Registry: Genetic structure of IDDM1: two separate regions in the major histocompatibility complex contribute to susceptibility or protection. Diabetes 47:263 269, 1998 12. Lie BA, Todd JA, Pociot F, Nerup J, Akselsen HE, Joner G, Dahl Jorgensen K, Ronningen KS, Thorsby E, Undlien DE: The predisposition to type 1 diabetes linked to the human leukocyte antigen complex includes at least one non class II gene. Am J Hum Genet 64:793 800, 1999 13. Hattori M, Yamato E, Itoh N, Senpuku H, Fujisawa T, Yoshino M, Fukuda M, Matsumoto E, Toyonaga T, Nakagawa I, Petruzzelli M, McMurray A, Weiner H, Sagai T, Moriwaki K, Shiroishi T, Maron R, Lund T: Cutting edge: homologous recombination of the MHC class I K region defines new MHC-linked diabetogenic susceptibility gene(s) in nonobese diabetic mice. J Immunol 163: 1721 1724, 1999 14. The MHC Sequencing Consortium: Complete sequence and gene map of a human major histocompatibility complex. Nature 401:921 923, 1999 15. Mizuki N, Ota M, Kimura M, Ohno S, Ando H, Katsuyama Y, Yamazaki M, Watanabe K, Goto K, Nakamura S, Bahram S, Inoko H: Triplet repeat polymorphism in the transmembrane region of the MICA gene: a strong association of six GCT repetitions with Behcet disease. Proc Natl Acad Sci U S A 94:1298 1303, 1997 16. Degli Esposti MA, Andreas A, Christiansen FT, Schalke B, Albert E, Dawkins RL: An approach to the localization of the susceptibility genes for generalized myasthenia gravis by mapping recombinant ancestral haplotypes. Immunogenetics 35:355 364, 1992 17. Gambelunghe G, Falorni A, Ghaderi M, Laureti S, Tortoioli C, Santeusanio F, Brunetti P, Sanjeevi CB: Microsatellite polymorphism of the MHC class I chain-related (MIC-A and MIC-B) genes marks the risk for autoimmune Addison s disease. J Clin Endocrinol Metab 84:3701 3707, 1999 18. Bunce M, O Neill CM, Barnardo MC, Krausa P, Browning MJ, Morris PJ, Welsh KI: Phototyping: comprehensive DNA typing for HLA-A, B, C, DRB1, DRB3, DRB4, DRB5 & DQB1 by PCR with 144 primer mixes utilizing sequence-specific primers (PCR-SSP). Tissue Antigens 46:355 367, 1995 19. Foissac A, Cambon-Thomsen A: Microsatellites in the HLA region: 1998 update. Tissue Antigens 52:318 352, 1998 20. Raymond M, Rousset F: An exact test for population differentiation. Evolution 49:1280 1283, 1995 DIABETES, VOL. 49, DECEMBER 2000 2221